Section_25
iCX Member
- Total Posts : 333
- Reward points : 0
- Joined: 2007/10/18 18:41:00
- Location: PA, USA
- Status: offline
- Ribbons : 2
I had my CPU just under 100% load at a max core temp of 65-73c for around 5 hours the other night and it got me wondering how long CPU's in general are supposed to or are estimated to run at or around those temps under full load before they fail.
Surely at 100% load after so long something would downclock or fail or something. Even with adequate cooling the chip would die eventually right ?
anyone know ?
CPU: Intel Core i7 CPU 920 @ 2.6 GHz (Stock) Motherboard: ASUS Sabertooth x58 RAM: G.SKILL Ripjaws (6GB) DDR3 GPU: GTX 770 SSD: Kingston 120gb PSU: Corsair TX750W Heatsink: Corsair H50 (Push/Pull) Display: Acer 24" Widescreen 1920x1080 O/S: Windows 8.1 Pro 64x
|
Flint 1760
Omnipotent Enthusiast
- Total Posts : 8295
- Reward points : 0
- Joined: 2009/04/26 15:44:26
- Status: offline
- Ribbons : 45
Re:How Long Before Failure ?
2010/04/03 13:20:17
(permalink)
Don't know for sure, however it would seem logical that it would shorten the service life as heat is a killer of electronics and you are pumping some watts through the chip.
That being said, the length of time may be after years and it is in drawer rattling around with other long forgotten hardware.
|
Section_25
iCX Member
- Total Posts : 333
- Reward points : 0
- Joined: 2007/10/18 18:41:00
- Location: PA, USA
- Status: offline
- Ribbons : 2
Re:How Long Before Failure ?
2010/04/03 14:33:41
(permalink)
Let me re-phrase my question.
If you have a CPU that you run 24/7, that occasionaly runs at 90-100% load, but most of the time it idles at almost 0% load and a CPU that you have running 90-100% load 24/7 how much longer do you think the chip that idles most of the time will live ?
I know its impossible to give an exact time period, but I was just looking for an estimate from someone that runs chips that hard and might know.
CPU: Intel Core i7 CPU 920 @ 2.6 GHz (Stock) Motherboard: ASUS Sabertooth x58 RAM: G.SKILL Ripjaws (6GB) DDR3 GPU: GTX 770 SSD: Kingston 120gb PSU: Corsair TX750W Heatsink: Corsair H50 (Push/Pull) Display: Acer 24" Widescreen 1920x1080 O/S: Windows 8.1 Pro 64x
|
i_cruz
iCX Member
- Total Posts : 491
- Reward points : 0
- Joined: 2008/12/26 20:36:36
- Location: Central Ohio
- Status: offline
- Ribbons : 1
Re:How Long Before Failure ?
2010/04/03 14:38:18
(permalink)
That is a good question!
I fold with my rig so the CPU and GPU are both kept at about 99% load 24/7. Temps for the CPU and the GPU are about 78*C.
I really don't know for how long it will keep running but it has been OK for the past 3 months!
Cheers!
IC
> PSU: OCZ Fatal1ty 1000W> Mainboard: EVGA Z690 CLASSIFIED (BIOS 2.08)> AIO: EVGA CLC 280mm AIO> CPU: Intel i7-13700k 3.4GHz> Physical Memory: G.SKILL Trident Z5 64GB (4 x 16GB) DDR5 > Physical Memory Model#: F5-6400J3239G16GA2-TZ5RK> Monitor: Samsung G9> Video Card: EVGA RTX 3090 Ti FTW3[URL=https://www.evga.com/associates/default.aspx?associatecode=HV4ZYIDMNODLH84]  [
|
lehpron
Regular Guy
- Total Posts : 16254
- Reward points : 0
- Joined: 2006/05/18 15:22:06
- Status: offline
- Ribbons : 191
Re:How Long Before Failure ?
2010/04/03 17:06:47
(permalink)
Section_25
how long CPU's in general are supposed to or are estimated to run at or around those temps under full load before they fail. Intel warrants their CPUs for 24/7 non-stop (= never shut down) 100% use for 3yrs at stock speeds as a minimum, mainly as the template design for all their CPUs are server CPUs. They see the most strenuous use beyond any laptop or desktop CPU, the only difference is in specifications. This is how we are gauranteed high quality. Note, after warranty doesn't mean failure, Intel determines warranty based on how many CPUs are likely to fail piror to warranty such that they can afford to replace them for free and not take a revenue hit, it is a business/technical limit. Their CPUs aren't 100% perfect; but overclocking voids warranty so it doesn't matter to us. Since we don't know how long each CPU can last, we can use Intel's warranty minimum as a reference for all of them, that they will last at least that long at stock. When you overclock, the lifespan does drop, but is mainly depends on heat wattage ratios. If you overclock to a level where the wattage is double stock, you get half the lifespan; it is exponential. But note, that just means you get at least 18 months of non-stop 24/7 100% use, if you use the CPU 8hrs/day that becomes 4.5yrs. But, in order to maintain your CPU stabilty; your RAM, HDD, motherboard, PSU, etc all have to be in sync and stable too. So while the CPU could theoretically take the punishment, total system stability depends on the other junk in the system.  Not saying temperature isn't an issue, but then Intel does not give out a range of acceptable temperatures for anything but stock use. Technically, the full load overclock needs to be less than what is stable for stock to maintain the longest life, it doesn't make sense to insist on the same temperature range (people here will say stay below 85C as this is what Intel's site states, I think this is wrong, and don't let my overclock pass 70C if I can help it). A simple formula for estimating the ratio of power due to an OC is (the Bclock ratio) x (multiplier ratio) x (CPU Vcore ratio) 2, ratios with respect to stock values (which is batch dependant). For example, my 920 C0 is set at 190x20 at Vcore of 1.4 from stock 1.3. So my wattage ratio is (190/133) x (20/20) x (1.4/1.3) 2 = 1.657, or I've cut my CPUs lifespan by 66% running it at 3.8GHz. Note, this assumes full load use, I very easily don't use all eight threads 100% every day, and mostly use it 6hr/day. I have much more than another 3yrs left and I've probably keep it that long. But my other components could fail before that, and affect the CPU, so just make sure your other components are top quality and you should be set for years. I wouldn't worry about a couple of months of use (or below 4Ghz), unless your CPU is experiencing an 80% overclock or in the range of 5GHz...
post edited by lehpron - 2010/04/03 17:17:13
|
Section_25
iCX Member
- Total Posts : 333
- Reward points : 0
- Joined: 2007/10/18 18:41:00
- Location: PA, USA
- Status: offline
- Ribbons : 2
Re:How Long Before Failure ?
2010/04/03 17:18:35
(permalink)
That is EXACTLY the type of response I was looking for lehpron..... thanks
Anyone else have anything to add ? lol
CPU: Intel Core i7 CPU 920 @ 2.6 GHz (Stock) Motherboard: ASUS Sabertooth x58 RAM: G.SKILL Ripjaws (6GB) DDR3 GPU: GTX 770 SSD: Kingston 120gb PSU: Corsair TX750W Heatsink: Corsair H50 (Push/Pull) Display: Acer 24" Widescreen 1920x1080 O/S: Windows 8.1 Pro 64x
|
Section_25
iCX Member
- Total Posts : 333
- Reward points : 0
- Joined: 2007/10/18 18:41:00
- Location: PA, USA
- Status: offline
- Ribbons : 2
Re:How Long Before Failure ?
2010/04/03 19:46:22
(permalink)
I certainly don't want to discourage anyone to participate in the folding program, but if this could potentialy damage hardware, I think it should be brought to everyones attention.
CPU: Intel Core i7 CPU 920 @ 2.6 GHz (Stock) Motherboard: ASUS Sabertooth x58 RAM: G.SKILL Ripjaws (6GB) DDR3 GPU: GTX 770 SSD: Kingston 120gb PSU: Corsair TX750W Heatsink: Corsair H50 (Push/Pull) Display: Acer 24" Widescreen 1920x1080 O/S: Windows 8.1 Pro 64x
|
lehpron
Regular Guy
- Total Posts : 16254
- Reward points : 0
- Joined: 2006/05/18 15:22:06
- Status: offline
- Ribbons : 191
Re:How Long Before Failure ?
2010/04/03 22:27:00
(permalink)
There's a saying (I'm paraphrasing) among the overclocking community, that "if you can't afford to replace, then you probably shouldn't be doing anything". Many here will replace long before a problem can occur due to a lifespan issue of any component. But when it happen [early due to a low quality batch, for example], many willl BMS for a while (blame the brand, blame the CS/TS, blame the shipper, etc), but eventually replace it [again]. The awesome headroom we see from overclocking is just a byproduct of design, these things are simply meant to take major stress loads at stock speeds for years of endurance more than any gamer/enthusiast is willing to keep. We can't act like some girl that thinks just because a pair of jeans fits her so well it must have been made for her; Intel does not make these CPUs for us, thus no reason to be a fan.  Regarding folding, I think ultimately a Xeon/Opteron would be the best option simply as they would be in its element, used as meant to be. But since many of us overclock and there aren't many multi-CPU boards with the option, we choose desktop CPUs.
|
SLeePYG72786
Superclocked Member
- Total Posts : 233
- Reward points : 0
- Joined: 2009/11/20 05:06:17
- Status: offline
- Ribbons : 0
Re:How Long Before Failure ?
2010/04/06 14:45:03
(permalink)
|
codeman
iCX Member
- Total Posts : 275
- Reward points : 0
- Joined: 2010/10/25 19:58:33
- Location: Toronto, Canada
- Status: offline
- Ribbons : 0
Re:How Long Before Failure ?
2011/02/17 19:36:25
(permalink)
Personally - why buy these and not use them the most and hardest you can? The longevity of the CPUs in the average and even extreme usage scenarios in home desktops will never kill these things unless you go to absolute extremes.
I have CPUs that work, that were overlocked to their maximum point of stability from 1996 that still work today. Especially the older CPU's where the nodes were much larger, there's less room for damage - but even the latest and greatest CPUs... it takes A LOT of heat and energy to fry a piece of processed sand. You're more likely to burn out or short a pin in a CPU socket from overvoltage, or pop a capacitor than you are to fry a CPU as long as it is kept within thermal specifications.
Intel i7 980x @ 4.48Ghz || 12GB Mushkin Blackline || Gigabyte X58A-UD5 EVGA GTX580 with HeatKiller WB @ 900/2300 with 1.14 volts 2x 120GB Mushkin Callisto Deluxe SSD's in RAID0 1.5TB WD Green storage drive (excellent quiet drive with optimized vibration) Xonar Essence ST to Harmon Kardon AMP w/Klipsch floor monitors CaseLabs M* Enermax Modu87+ 800watt PSU 27" Apple LED Cinema Display Windows 7 x64 Ultimate Mod Rig
|
emach
FTW Member
- Total Posts : 1106
- Reward points : 0
- Joined: 2009/05/25 00:03:36
- Location: Park Ridge, IL
- Status: offline
- Ribbons : 6
Re:How Long Before Failure ?
2011/02/17 20:34:22
(permalink)
lehpron
Section_25
how long CPU's in general are supposed to or are estimated to run at or around those temps under full load before they fail.
Intel warrants their CPUs for 24/7 non-stop (= never shut down) 100% use for 3yrs at stock speeds as a minimum, mainly as the template design for all their CPUs are server CPUs. They see the most strenuous use beyond any laptop or desktop CPU, the only difference is in specifications. This is how we are gauranteed high quality. Note, after warranty doesn't mean failure, Intel determines warranty based on how many CPUs are likely to fail piror to warranty such that they can afford to replace them for free and not take a revenue hit, it is a business/technical limit. Their CPUs aren't 100% perfect; but overclocking voids warranty so it doesn't matter to us.
Since we don't know how long each CPU can last, we can use Intel's warranty minimum as a reference for all of them, that they will last at least that long at stock. When you overclock, the lifespan does drop, but is mainly depends on heat wattage ratios. If you overclock to a level where the wattage is double stock, you get half the lifespan; it is exponential. But note, that just means you get at least 18 months of non-stop 24/7 100% use, if you use the CPU 8hrs/day that becomes 4.5yrs. But, in order to maintain your CPU stabilty; your RAM, HDD, motherboard, PSU, etc all have to be in sync and stable too. So while the CPU could theoretically take the punishment, total system stability depends on the other junk in the system.
Not saying temperature isn't an issue, but then Intel does not give out a range of acceptable temperatures for anything but stock use. Technically, the full load overclock needs to be less than what is stable for stock to maintain the longest life, it doesn't make sense to insist on the same temperature range (people here will say stay below 85C as this is what Intel's site states, I think this is wrong, and don't let my overclock pass 70C if I can help it).
A simple formula for estimating the ratio of power due to an OC is (the Bclock ratio) x (multiplier ratio) x (CPU Vcore ratio)2, ratios with respect to stock values (which is batch dependant).
For example, my 920 C0 is set at 190x20 at Vcore of 1.4 from stock 1.3. So my wattage ratio is (190/133) x (20/20) x (1.4/1.3)2 = 1.657, or I've cut my CPUs lifespan by 66% running it at 3.8GHz. Note, this assumes full load use, I very easily don't use all eight threads 100% every day, and mostly use it 6hr/day. I have much more than another 3yrs left and I've probably keep it that long. But my other components could fail before that, and affect the CPU, so just make sure your other components are top quality and you should be set for years.
I wouldn't worry about a couple of months of use (or below 4Ghz), unless your CPU is experiencing an 80% overclock or in the range of 5GHz...
I wonder if this holds true for the 1155 platform as well, just seems that people are reaching higher overclocks than ever before on consumer level cooling solutions. The less voltage req and lower heat production of these cpu's at higher overclocks should improve life expectancy, should it not?
|
lehpron
Regular Guy
- Total Posts : 16254
- Reward points : 0
- Joined: 2006/05/18 15:22:06
- Status: offline
- Ribbons : 191
Re:How Long Before Failure ?
2011/02/20 00:43:09
(permalink)
Mathematically, architecture is irrelevant, all that matters is the ratio of wattages; so yeah it would apply to Sandy Bridge. But, Intel gives the same minimum warranty condition to all their processors, so having a higher stock frequency and not needing as much votlage to go higher isn't compariable towards an older CPU that did require more voltage for higher speeds.
For example, a C2D E8600 would not have needed as much of an increase in voltage for 4GHz as an E8200 because of their stock frequencies. Also despite similar TDP's, the actual heat loads are different, it is the same with any CPU of the same TDP range. I think it is posible folks are misreading the high OC'ing of the SB processors simply due to the high stock speeds. If a 2.66GHz part debuted first, the picture would be entirely different, and require more voltage changes.
Always compare to [your specific] stock, it does no good to compare to an entirely different processor type.
|
deegon
New Member
- Total Posts : 3
- Reward points : 0
- Joined: 2010/10/21 02:38:25
- Status: offline
- Ribbons : 0
Re:How Long Before Failure ?
2011/04/21 18:48:55
(permalink)
I'm with the belief that one should run with what the stock temps are when the stock cooler is in use and that can be from 55Deg.c all the way to 85Deg.c for my chip u must remember that u are over clocking and allot of the thoughts on the subject are points of view, the best thing to do is to ask as many people as u can and go from there, because it is all about the "i think". the faster u go the bigger the mess
post edited by deegon - 2011/04/22 05:28:02
|
rafale
FTW Member
- Total Posts : 1554
- Reward points : 0
- Joined: 2010/01/09 20:11:47
- Location: Portland, OR
- Status: offline
- Ribbons : 6
Re:How Long Before Failure ?
2011/05/03 16:42:19
(permalink)
lehpron
Section_25
how long CPU's in general are supposed to or are estimated to run at or around those temps under full load before they fail.
Intel warrants their CPUs for 24/7 non-stop (= never shut down) 100% use for 3yrs at stock speeds as a minimum, mainly as the template design for all their CPUs are server CPUs. They see the most strenuous use beyond any laptop or desktop CPU, the only difference is in specifications. This is how we are gauranteed high quality. Note, after warranty doesn't mean failure, Intel determines warranty based on how many CPUs are likely to fail piror to warranty such that they can afford to replace them for free and not take a revenue hit, it is a business/technical limit. Their CPUs aren't 100% perfect; but overclocking voids warranty so it doesn't matter to us.
Since we don't know how long each CPU can last, we can use Intel's warranty minimum as a reference for all of them, that they will last at least that long at stock. When you overclock, the lifespan does drop, but is mainly depends on heat wattage ratios. If you overclock to a level where the wattage is double stock, you get half the lifespan; it is exponential. But note, that just means you get at least 18 months of non-stop 24/7 100% use, if you use the CPU 8hrs/day that becomes 4.5yrs. But, in order to maintain your CPU stabilty; your RAM, HDD, motherboard, PSU, etc all have to be in sync and stable too. So while the CPU could theoretically take the punishment, total system stability depends on the other junk in the system.
Not saying temperature isn't an issue, but then Intel does not give out a range of acceptable temperatures for anything but stock use. Technically, the full load overclock needs to be less than what is stable for stock to maintain the longest life, it doesn't make sense to insist on the same temperature range (people here will say stay below 85C as this is what Intel's site states, I think this is wrong, and don't let my overclock pass 70C if I can help it).
A simple formula for estimating the ratio of power due to an OC is (the Bclock ratio) x (multiplier ratio) x (CPU Vcore ratio)2, ratios with respect to stock values (which is batch dependant).
For example, my 920 C0 is set at 190x20 at Vcore of 1.4 from stock 1.3. So my wattage ratio is (190/133) x (20/20) x (1.4/1.3)2 = 1.657, or I've cut my CPUs lifespan by 66% running it at 3.8GHz. Note, this assumes full load use, I very easily don't use all eight threads 100% every day, and mostly use it 6hr/day. I have much more than another 3yrs left and I've probably keep it that long. But my other components could fail before that, and affect the CPU, so just make sure your other components are top quality and you should be set for years.
I wouldn't worry about a couple of months of use (or below 4Ghz), unless your CPU is experiencing an 80% overclock or in the range of 5GHz...
Wow, I just got to this post through a link in Lephron's sig and I am really puzzled by his claims. My first reaction was actually "what a bunch of crap!" I am not saying that they are wrong but I just don't know where and how he came up with this math. As a semiconductor engineer working in the manufacturing process of these chips and dealing with defects and reliability issues, even I would not go as far as making these claims. If they are empirical statistics it is fine but I can tell you that they have no technical or theoritical foundations. A chip can fail for several reasons and at multiple levels from the front end device level to the interconnect and all the way to the packaging side and every design rule and technology node will behave differently. It is impossible to come up with a formulas like what Lephron is proposing to predict failure and even less that wattage ratio absurdity. Over the years I have seen so many mechanisms with completely different behaviors and effects over lifetime that I really know what to think of something like this. Although I don't disagree with Lephron's philosophy and his conclusions on overclocking that it will reduce the lifetime of the CPU but is probably not something that should stop us from doing it because the hardware will probably obsolete by the time we risk seeing a problem at the rate we use them, the pseuso science behind it is just mind boggling.
post edited by rafale - 2011/05/03 16:44:58
|
slickric
CLASSIFIED Member
- Total Posts : 3401
- Reward points : 0
- Joined: 2007/07/18 19:27:16
- Location: Hollywood, Florida USA
- Status: offline
- Ribbons : 17
Re:How Long Before Failure ?
2011/05/04 00:18:30
(permalink)
Section_25
I had my CPU just under 100% load at a max core temp of 65-73c for around 5 hours the other night and it got me wondering how long CPU's in general are supposed to or are estimated to run at or around those temps under full load before they fail.
Surely at 100% load after so long something would downclock or fail or something. Even with adequate cooling the chip would die eventually right ?
anyone know ?
I don't know how long CPU's last under overclocking but mine has ben OC'ed for a couple of years. My ambient temps are 22 to 24 degrees.. I am in South Florida but run AC all the time and am OC'ed all the time. I game all the time...
|
_Nite_
Insert Custom Title Here
- Total Posts : 19175
- Reward points : 0
- Joined: 2007/04/18 15:35:04
- Location: Sumner, Washington
- Status: offline
- Ribbons : 46
Re:How Long Before Failure ?
2011/05/04 03:35:55
(permalink)
I use the stock cooler on my i7 920 @ 2.8Ghz with the cores hovering around the 80c range on 100% load 24/7 with folding@home whenever I'm not gaming, and its been like that for 2 years solid now and still running great
I did my build in april 2009 and here it is May 2011
Gaming Rig: AMD Ryzen 7 2700x @ 4.25Ghz - Asrock B450 Pro 4 - 16GB Crucial Ballistix DDR4 3000 - 240GB SSD (OS), 1TB ADATA SU800 SSD and 2TB ADATA SX8200 Pro M2 NVME (Games) - 8GB RTX 3070 XC3 Ultra - Sound Blaster Audigy 5 RX - EVGA Supernova 750w G2 - Antec 302 - 24" Asus VG248QE
|
drnilly007
Superclocked Member
- Total Posts : 106
- Reward points : 0
- Joined: 2006/04/12 10:08:31
- Status: offline
- Ribbons : 1
Re:How Long Before Failure ?
2011/05/11 17:44:28
(permalink)
The best way to find out is to try and run it into the ground LOL. JK dont try that
the Way, the Truth, the Life...did you know your His favorite!
|
feniks
CLASSIFIED ULTRA Member
- Total Posts : 7983
- Reward points : 0
- Joined: 2008/08/02 22:07:39
- Location: NYC
- Status: offline
- Ribbons : 38
Re:How Long Before Failure ?
2011/06/03 22:03:16
(permalink)
it depends what you guys define as "chip fails". surely it will slowly start degrading when 100% load is kept at max temps around 70-80C ... but such degradation can actually take months to affect the stability of chip, no kidding. and usually it takes a little more core voltage to make it stable again (with a little higher temps) and so on.
even after 2-3 years of such use of chip you will find that it is still perfectly stable at stock speeds and stock voltage with quite normal temps :)
ask me how I now :)
|
Xtrem
FTW Member
- Total Posts : 1957
- Reward points : 0
- Joined: 2010/02/12 20:49:09
- Location: Los Angeles
- Status: offline
- Ribbons : 5
Re:How Long Before Failure ?
2011/06/04 04:05:19
(permalink)
Its weird, I haven't heard any CPU's die yet. Has anyone?
Order through banner below for 5-10% off your order at EVGA.com  Register your EVGA products using FKDP4PLDDB Elite Referral
|
dogofwar
SSC Member
- Total Posts : 599
- Reward points : 0
- Joined: 2008/04/13 20:45:02
- Status: offline
- Ribbons : 1
Re:How Long Before Failure ?
2011/06/04 08:48:13
(permalink)
emach
lehpron
Section_25
how long CPU's in general are supposed to or are estimated to run at or around those temps under full load before they fail.
Intel warrants their CPUs for 24/7 non-stop (= never shut down) 100% use for 3yrs at stock speeds as a minimum, mainly as the template design for all their CPUs are server CPUs. They see the most strenuous use beyond any laptop or desktop CPU, the only difference is in specifications. This is how we are gauranteed high quality. Note, after warranty doesn't mean failure, Intel determines warranty based on how many CPUs are likely to fail piror to warranty such that they can afford to replace them for free and not take a revenue hit, it is a business/technical limit. Their CPUs aren't 100% perfect; but overclocking voids warranty so it doesn't matter to us.
Since we don't know how long each CPU can last, we can use Intel's warranty minimum as a reference for all of them, that they will last at least that long at stock. When you overclock, the lifespan does drop, but is mainly depends on heat wattage ratios. If you overclock to a level where the wattage is double stock, you get half the lifespan; it is exponential. But note, that just means you get at least 18 months of non-stop 24/7 100% use, if you use the CPU 8hrs/day that becomes 4.5yrs. But, in order to maintain your CPU stabilty; your RAM, HDD, motherboard, PSU, etc all have to be in sync and stable too. So while the CPU could theoretically take the punishment, total system stability depends on the other junk in the system.
Not saying temperature isn't an issue, but then Intel does not give out a range of acceptable temperatures for anything but stock use. Technically, the full load overclock needs to be less than what is stable for stock to maintain the longest life, it doesn't make sense to insist on the same temperature range (people here will say stay below 85C as this is what Intel's site states, I think this is wrong, and don't let my overclock pass 70C if I can help it).
A simple formula for estimating the ratio of power due to an OC is (the Bclock ratio) x (multiplier ratio) x (CPU Vcore ratio)2, ratios with respect to stock values (which is batch dependant).
For example, my 920 C0 is set at 190x20 at Vcore of 1.4 from stock 1.3. So my wattage ratio is (190/133) x (20/20) x (1.4/1.3)2 = 1.657, or I've cut my CPUs lifespan by 66% running it at 3.8GHz. Note, this assumes full load use, I very easily don't use all eight threads 100% every day, and mostly use it 6hr/day. I have much more than another 3yrs left and I've probably keep it that long. But my other components could fail before that, and affect the CPU, so just make sure your other components are top quality and you should be set for years.
I wouldn't worry about a couple of months of use (or below 4Ghz), unless your CPU is experiencing an 80% overclock or in the range of 5GHz...
I wonder if this holds true for the 1155 platform as well, just seems that people are reaching higher overclocks than ever before on consumer level cooling solutions. The less voltage req and lower heat production of these cpu's at higher overclocks should improve life expectancy, should it not?
this is what technology is all about!-STRONGER--BETTER--FASTER.. 
ASRock Z790 Steel Legend, DG-75 case INTEL I7 13700K LGA1700 @ stock clock MSI Geforce RTX 3060-TICorsair Vengeance DDR5 32 GB EVGA 1000 G5 PSUEVGA CLCx240970 EVo M.2 Win10 64EVGA CLC 240x120 WIN7/64, ANTEC 1200, H60 CPU COOLER, CORSAIR SSD SAT3
|
lehpron
Regular Guy
- Total Posts : 16254
- Reward points : 0
- Joined: 2006/05/18 15:22:06
- Status: offline
- Ribbons : 191
Re:How Long Before Failure ?
2011/06/04 10:55:36
(permalink)
It is my belief that the LGA1155 quads have way less starting stock wattages (due to process tech) to the point where they behave like lower-cored CPUs as we overclock -- if Intel had a business case a 6-core could fit in 95W TDP. Additionally, the higher starting frequency of the LGA1155 quads can imply higher overclocks when its just the same ratio we're used to seeing: from 2.66GHz to 4GHz is 50%, from 3.4GHz to 5.1GHz is also 50%. Some are just superficial about numbers... As for hearing about dead CPUs, I'm more inclinded to think 1) it is embarrassing to admit burning them out from overclocking and 2) with respect to desktop PC uses, Intel's parts are overdesigned since the 3yr 24/7 use warranty would stretch out to 9yrs at 8hr/day, while overclocked 50% becomes 4yrs of 8hrs/day -- and that it just to the reference of warranty period. Note: After 3 years, Intel stock parts don't just die, warranties are decided based on a material/manufacturing fatigue limit of the processor where after so many cycles of stress over a period of time, a small percentage of processors are expected to fail. Intel determines how many they can afford to replace for free via RMA's and still make a buck on the rest of them. Meaning a vast majority of Intel parts at stock will last longer than 3yrs even used at 24/7 nonstop -- as such our parts overclocked by 50% may last longer than 4yrs if used 8hr/day -- but how many of us will keep an overclocked part that long? That's why we won't hear about CPU's dying as often.
|
feniks
CLASSIFIED ULTRA Member
- Total Posts : 7983
- Reward points : 0
- Joined: 2008/08/02 22:07:39
- Location: NYC
- Status: offline
- Ribbons : 38
Re:How Long Before Failure ?
2011/06/04 11:37:48
(permalink)
^ exactly what lehpron says. I kept my Q9450 overclocked 24/7 (I treat my rig as a server and gaming system and multimedia system all in one) from stock 2.66GHz to 3.8GHz for months. then due to temps (while folding) I lowered the clock to 3.6GHz for another few months. then I loaded it with 4 sticks memory, 3 video cards and what not and used max stable overclock at 3.4GHz (limited only by North Bridge capabilities on my board) for 1.5years (24/7) or so. now again (still 24/7) at 3.6GHz and running just fine. of course now it takes a little more voltage (roughly 0.04V difference as set in BIOS) to be same stable at 3.6GHz with "diet" setup (only 2 memory sticks and 2 video cards as North Bridge on my board is not too awesome). All overclocks were confirmed with a fearful (not to me though) Intel Burn Test, 20 rounds at max stress where CPU heats up really well .. and I do not use water cooling, only air cooling and having some hot ambient room temps (apartment right below roof, from sunny side).
post edited by feniks - 2011/06/04 22:57:12
|
Xtrem
FTW Member
- Total Posts : 1957
- Reward points : 0
- Joined: 2010/02/12 20:49:09
- Location: Los Angeles
- Status: offline
- Ribbons : 5
Re:How Long Before Failure ?
2011/06/04 15:23:32
(permalink)
So running 50% running 24/7 would mean 1.33 years!?
Order through banner below for 5-10% off your order at EVGA.com Register your EVGA products using FKDP4PLDDB Elite Referral
|
feniks
CLASSIFIED ULTRA Member
- Total Posts : 7983
- Reward points : 0
- Joined: 2008/08/02 22:07:39
- Location: NYC
- Status: offline
- Ribbons : 38
Re:How Long Before Failure ?
2011/06/04 16:08:53
(permalink)
yeah, something like that. not all chips are same, also different batches (based on production dates) are different too. Xtrem
So running 50% running 24/7 would mean 1.33 years!?
|
lehpron
Regular Guy
- Total Posts : 16254
- Reward points : 0
- Joined: 2006/05/18 15:22:06
- Status: offline
- Ribbons : 191
Re:How Long Before Failure ?
2011/06/05 10:10:05
(permalink)
Xtrem
So running 50% running 24/7 would mean 1.33 years!? Mathematically, the inverse of the wattage ratio is the lifespan ratio, but like feniks says, the batch will vary what clockspeed determines your wattage change. Just mess around with the formula in my earlier post to get an estimate based on your CPU.
|
deegon
New Member
- Total Posts : 3
- Reward points : 0
- Joined: 2010/10/21 02:38:25
- Status: offline
- Ribbons : 0
Re:How Long Before Failure ?
2011/06/09 02:30:57
(permalink)
As with many devices, cars, phones, double glazed windows and so on, manufactures will give warranties and guaranty on products. As not all products are perfect when made and this gives a time period in which break downs or failures can occur if any. As 99.99% of products out live the given Warranties many times over it is not uncommon to have CPUs out live every other part of a computer, it is more likely that a failing motherboard or power supply will damage or destroy the CPU than out living the true useable life of the component. In most cases the computer after 10years will be obsolete  I have siting next to me a computer the is about 10+ years old and is now too slow and power hungry to run a full blast and has been run at over clock speeds all its life, i have had many computers that have failed due to motherboard failures although CPU fail does happen Guaranty *something that assures a particular outcome or condition Warrants * a written guarantee given to the purchaser of a new appliance, automobile, or other item by the manufacturer or dealer, usually specifying that the manufacturer will make any repairs or replace defective parts free of charge for a stated period of time.
post edited by deegon - 2011/06/09 03:27:33
|
rafale
FTW Member
- Total Posts : 1554
- Reward points : 0
- Joined: 2010/01/09 20:11:47
- Location: Portland, OR
- Status: offline
- Ribbons : 6
Re:How Long Before Failure ?
2011/06/09 12:04:49
(permalink)
The warranty has indeed nothing to do with the life expectancy of the product. The chips are tested and specified with some models which may or may not be correct. I honestly don't know. DRAM for example is built with so much redundancy that the most likely failure is the transistor itself. They run their stress test at 60 or 80C! Logic chips like CPU, GPU, NB, SB are much more complex and run their test up to 200C for up to one month. The number of good transistor needs to exceed a certain percentage to call the manufacturing process good. Obviously they test samples and not every chip since the chip is dead by the end of the test.
I can only tell you that the failure rates are not a linear function of temperature or voltage or current. They tend to be exponential curves or at least parabolic depending on the factor and the failure mode but when the sudden ramp of failure start really depends on the chip design and manufacturing process. You cannot use some random equation to model them. This is just BS.
What is true is that the working window for current and temperature defined by manufacturers are usually in the very safe zone with some margin before the exponential increase of failure in the test which is what enables us to overclock and overvolt. (although sometimes, for some of the recent GPUs I really wonder how far they are.) The product will most likely be obsolete by the time you will see a failure and if early failure occur, it could be because the particular chip had some defect. This is why there is a warranty...
|
funk695
New Member
- Total Posts : 26
- Reward points : 0
- Joined: 2009/02/12 11:41:06
- Status: offline
- Ribbons : 0
Re:How Long Before Failure ?
2011/08/03 13:56:47
(permalink)
i7 920 co stepping at 4.2 1.35v w/ vdrop on for almost 3 years now. No degradation in performance yet and i live in the dessert.
|
shaneduce
CLASSIFIED Member
- Total Posts : 2017
- Reward points : 0
- Joined: 2009/10/21 10:19:14
- Location: Portland ,Or.
- Status: offline
- Ribbons : 3
Re:How Long Before Failure ?
2011/08/06 23:29:55
(permalink)
Xtrem
Its weird, I haven't heard any CPU's die yet. Has anyone?
Nope, and i've had every thing die but the CPU. I have a old Northwood 478 cpu had it clocked at 4 ghz 16 mult x 250 fsb. I had it runing like that for 5+ years. That CPU was in a set up that the PSU spicked the MB and set it on fire. The only thing that survied it was the that CPU
|
jokmi
New Member
- Total Posts : 3
- Reward points : 0
- Joined: 2010/12/22 04:34:41
- Status: offline
- Ribbons : 0
Re:How Long Before Failure ?
2011/09/03 01:53:50
(permalink)
Yep, my 486 (about 100 MHz) Intel proc died 1995..7. Replaced with 486 amd (about 100..120 MHz). Crappy proc fan had failed and I guess the proc wasn't fast enough to stop its things under over-temp. The amd worked for years until p2 350, which was my first oc (to 410..430 (?)). Should've bought Celly300 (way cheap 450 oc). Many OCs since then with no broken cpus.
Wondering how much better/faster, if at all, are today's cpus at lowering wattage when tripped to overtemp. Too lazy to dig the info from Intel's or Google's.
|